Choose a process that takes more effort than it should. I built this tool to help you map it out, spot what’s slowing it down, and work with your AI on a better way to get it done.
Give yourself around 30 minutes. You’ll leave with your workflow mapped and a prompt to start working through improvements with your AI.
Your Mapper is ready.
Confirm your email to open it. This browser will remember your access.
You’re ready to map. This browser remembers your access.
See where AI fits in my LinkedIn workflow.
The steps, the decisions and what passes from one stage to the next.
Your download contains the editable workflow and a prompt for creating a diagram in your AI. The Mapper does not generate the image itself. View full size ↗
Step 1 of 9
Choose one workflow.
Pick a contained piece of work and follow it in detail. “Create LinkedIn posts” is broad. A useful map shows where ideas come from, how briefs and drafts are created, who reviews them, where files go and what happens next.
Is my workflow specific enough?
Choose a repeated result, not a department. “Client onboarding” can work. “Operations” is too broad.
Step 2 of 9
Choose the right route.
Are you mapping how the work happens today, or a first workable version of something new?
Step 3 of 9
Set the boundary.
Make the beginning and end precise so the workflow cannot quietly expand.
Use one event, deadline or arrival. A pool of ideas or source material belongs in the next step.
Finish with something delivered, approved, stored or scheduled. “Generating interest” is a business outcome, not the end of the workflow.
This reads like an ongoing business outcome. Name the deliverable or handoff that ends this workflow, such as approved posts being scheduled.
Step 4 of 9
Collect the inputs.
List what must be available before the work can be done properly.
Think about forms, transcripts, examples, policies, software access and approvals.
Step 5 of 9
Break the workflow into stages.
Follow the work from the trigger to the finish. Start with one stage, then add and reorder stages as the process becomes clearer.
One checkpoint = one result along the way. For example: “The intake material is complete”, then “Missing information is flagged”, then “The draft brief is ready for review”.
How should I write each stage?
Describe one result that becomes true before the workflow can move forward. Write “Missing information is flagged” rather than “Read the transcript.”
Step 6 of 9
Decide where human judgment must stay.
Imagine the workflow moving without you. At what point must it stop, change direction or ask a person to decide?
Look for the moments where getting it wrong matters. Client suitability, pricing, creative direction, brand voice, legal commitments and anything sent externally usually need a clear human decision or approval.
Step 7 of 9
Find the friction.
Mark where work slows down, creates rework or returns to the owner.
Step 8 of 9
Choose one AI responsibility to test first.
Your map may reveal several opportunities. Pick one contained responsibility as the starting point, then capture the others for later.
Stop here before building. A dependable teammate still needs a role, instructions, approved knowledge and testing. This map gives you the foundation.
Step 9 of 9
Your workflow map.
Review the result. You can return to any step to make changes before exporting.
Export your map
Download a clean, editable copy of your completed workflow.
Continue later
Save an editable progress file, then load it again if you want to keep working.
Visualize it with AI
Copy a prompt that includes this workflow map, then paste it into ChatGPT, Claude or Gemini to create a standalone HTML visual.
Preview the prompt
Your workflow is mapped
Now build the AI agents that can help with the work.
Choose the course and live workshops if you want to build them yourself with my guidance. Choose private implementation if you want me to audit the wider business and build the agreed tools and agents with you.
Your workflow stays private. Your answers remain in this browser unless you download them. I only measure anonymous visits and button use.
Real client workflow
Rona's client inquiry-to-proposal workflow.
Rona's process began when a prospective client completed the inquiry form on her website. She then had to review the answers, arrange and lead the discovery conversation, shape the scope and prepare the proposal herself. Information sat across the form, notes and memory, so the entire path to a proposal depended on her.
The boundary
TriggerA prospective client submits the form on Rona’s website.
FinishThe prospect receives the proposed scope, quotation and next-step email.
Outside the workflowAccepting the proposal and formally onboarding the client begin the next workflow.
Current routeCapture the manual reality before deciding what AI should support.
How the work originally moved
1
The website form arrives by email
Rona opens it, reads the answers and decides whether the enquiry looks relevant. The original questions often provide only broad style preferences.
2
The prospect’s information is organised
A folder needs to be created and the submitted form, references and later notes need to be saved together.
3
A discovery conversation is arranged
Rona replies, handles the scheduling and prepares to ask for the information the form did not capture.
4
The discovery conversation is completed
Some calls are recorded and some happen in person. Rona relies heavily on notes and memory to preserve what matters.
5
The first brief and scope are formed
Rona interprets the form, conversation, constraints and visual preferences, then starts structuring the project herself.
6
The proposal package is prepared
She writes the scope, prepares the quotation and packages the information into an email or presentation.
7
The proposal and next step are sent
The prospect can accept, ask questions or decline. An accepted proposal moves into client onboarding.
What the map exposes
Manual pressureReading forms, creating folders, saving material, scheduling, taking notes, recalling the conversation, drafting the scope and writing the follow-up all depend on Rona.
Information riskThe form may not collect enough detail, recording is inconsistent and important context can remain in memory or scattered notes.
Possible AI starting pointRead the submitted form, organise the answers, flag missing information and prepare a short pre-call brief.
Human ownershipRona still decides fit, leads discovery, interprets the client and approves every scope, quotation and client-facing message.
This is where an AI team becomes practical.
The map shows several responsibilities that could stop depending entirely on Rona: preparing the pre-call brief, organising the transcript, drafting the project brief, packaging the proposal and preparing follow-up. She can now choose what to build first and give each AI role the right files, instructions and review points.
